Acid, P-deficient soils with a large capacity to immobilize P favour the rapid decomposition of apatite. Such soils are widespread in the east of South Africa and the possibility exists that Langebaan rock phosphate (Langfos) may be a more cost-effective source of P than superphosphate on these soils. This study compared the effectiveness of granular Langfos to that of double superphosphate in a pot experiment using a strongly P-fixing Balmoral clay (Typic Haplorthox) and a weakly P-fixing Avalon sandy loam (Plinthic Paleu- dult). Five rates of double superphosphate (0, 100, 250, 500 and 1 000 mg P kg−1) and six of Langfos (0, 100, 250, 500, 1 000 and 2 000 mg P kg−1) were compared at three lime rates. Three consecutive crops of maize (Zea maysL.) were grown for periods of four weeks. The average of the initial response slopes from the relationship between cumulative yield and P applied for the different lime applications showed that the effectiveness of Langfos relative to superphosphate was 10% and 18% for the Balmoral and Avalon soils, respectively. Lime had relatively little effect on plant response to superphosphate on the Balmoral, but had a marked positive effect on the Avalon. The response to Langfos was negatively affected by lime on the Balmoral, but only slightly so on the Avalon. These results suggest that granular Langfos will be more effective in acid sandy soils than in acid clay soils but that it is not an effective substitute in either for water-soluble P fertilizers.